1. Field of the Invention
The present invention relates to a fin field effect transistor (FET) and a method of manufacturing the same, and more particularly, relates to a double gate fin FET and a method of manufacturing the same.
2. Description of the Related Art
Complementary metal oxide silicon (CMOS) devices are widely used in various logic circuit applications such as central processing units and memory devices, and a great deal of added value can be generated in a CMOS-related industry. Accordingly, there has been intensive worldwide research has been conducted on CMOS devices based on nanotechnology. In many semiconductor device applications, the semiconductor device is required to exhibit low power consumption, small size and/or high operation speeds. CMOS devices are often capable of meeting all of these requirements.
However, as the size of CMOS devices continue to become reduced to very small levels, an effect known as a short channel effect can result as a consequence of the scaling-down of the gate size of the device. This short-channel effect can degrade the performance of the CMOS device.
Conventional CMOS devices are typically fabricated on bulk silicon substrates. However, in such conventional CMOS devices, the characteristics of the device may be strongly influenced by the processing conditions used in fabricating the device when the gate length of the device is scaled down to less than about 50 nm, and conventional CMOS devices may break down completely and cease to operate as semiconductor circuits when the gate length of the device is scaled down to less than about 30 nm.
In addition, although the gate length of the device is scaled down, the occupying area for a unit cell is not scaled down because a spacer on a sidewall of a gate electrode is not reduced, so that the size reduction of the gate length does not necessarily contribute to increase in the degree of integration of a semiconductor device.
The use of silicon-on-insulator (SOI) semiconductor devices (SOI devices) has been suggested as a way to provide CMOS devices with channel lengths of about 30 nm or less without the above technical limits of associated with bulk-silicon-based CMOS devices. However, fabrication of conventional CMOS devices on SOI substrates may result in an increase in the parasitic resistance at the source/drain regions because the silicon layer of the SOI substrate is very thin, and thus a selective epitaxial growth step to form elevated source/drain regions may be included when fabricating transistors in SOI devices. In addition, SOI devices may exhibit relatively poor contact with the substrate such that a floating body effect may arise and/or heat conduction between the device and the substrate may be reduced. All of these effects may operate to reduce device performance.
As described above, the SOI devices also do not necessarily contribute to reduction of a device size as compared with conventional bulk-silicon-based CMOS devices, so that the use of triple or double gate structures has been suggested as a way for efficiently reducing the channel length of CMOS devices to less than about 25 nm. Such CMOS devices are known as fin field effect transistors (fin FET).
When the channel is formed, both side and top surfaces of a fin, which is a pattern protruded from the bulk silicon substrate, a gate electrode is formed on each of the three surfaces of the fin, and a FET including three gate electrodes is known as a triple gate FET. In contrast, when a capping layer is formed on the top surface of the fin and a vertical gate field is interrupted by the capping layer, a gate electrode is formed on both side surfaces of the fin and a FET including two gate electrodes is known as a double gate FET.
Accordingly, since a gate electrode is formed on each surface of the channel region, fin FET transistors may provide improved gate control characteristics, which may result in reduced current leakage between source and drain electrodes of a transistor when compared with a conventional, single gate CMOS device, thereby improving a drain-induced barrier lowering (DIBL) phenomenon.
In addition, a threshold voltage of the fin FET may be varied dynamically because of the double gate; thus, a switching characteristic of a channel in the transistor may be considerably improved as compared with the conventional single gate. Therefore, a driving current of the transistor is increased and the short channel effect is sufficiently prevented.
However, the fin FET and the manufacturing method thereof have various associated problems. Such limitations will be described with reference to FIGS. 1-6. FIG. 1 is a perspective view illustrating a conventional triple gate fin FET, and FIG. 2 is a perspective view illustrating a conventional double gate fin FET. FIG. 3 is a perspective view illustrating a damaged capping layer of the double gate fin FET shown in FIG. 2.
Referring to FIG. 1, a width of a fin 20 in a triple gate fin FET 10 is reduced due to a recent trend of high integration in semiconductor devices; thus, an end portion A of the fin 10 becomes acute as various processes are repeated. Accordingly, an electric field that is applied in a vertical direction with respect to the fin 10 (hereinafter referred to as a vertical gate field), is focused into the end portion A of the fin 10. As a result, electric characteristics of the transistor are deteriorated due to the concentration of the vertical gate field, such as swing degradation below a threshold voltage, decrease of threshold voltage and increase of current leakage below the threshold voltage. Furthermore, the fin 20 is also etched away during an etching process against a gate conductive layer 30a for forming a gate electrode 30, so that a source/drain region 20a of the fin 20 is formed to be lower than a channel region 20b by as much as a distance B.
For the above reasons, the double gate fin FET 150 has been suggested as shown in FIG. 2. A capping layer 120 is formed on the fin 110 in the double gate fin FET 150. The capping layer 120 on the fin 110 controls the vertical gate field and maintains an original shape of the fin 110. Further, the capping layer 120 prevents a source/drain region of the fin 110 from being etched away when the fin 110 is under an etching process for forming a gate electrode 130.
However, the capping layer 120 is also unfavorably etched away during subsequent processes for forming a channel region, so that the original shape of the capping layer 120 is damaged and changed as shown in FIG. 3, or although not shown in FIG. 3, is completely removed from the fin 100. The damaged capping layer 120a also causes the same problem as known in the three gate fin FET.
Hereinafter, the reason for which the original shape of the capping layer is changed is described with reference to FIGS. 4 to 6.
FIGS. 4 to 6 are perspective views illustrating processing steps for a method of forming a conventional double gate fin FET.
Referring to FIG. 4, a silicon nitride layer is formed on a silicon substrate by a chemical vapor deposition (CVD) process. Then, the silicon nitride layer is patterned to thereby form a capping layer 120. The substrate is partially etched away using the capping layer 120 as an etching mask. Hereinafter, the etched substrate is designated as a reference numeral 100a in FIG. 4. The etching of the substrate forms a fin 110 protruded from the etched substrate 100a and extended along a first direction. A sidewall oxide layer 112 is formed on sidewalls of the fin 110 by a thermal oxidation process, so that damages on the sidewalls of the fin 110 are sufficiently cured. Then, a silicon nitride layer 114 is formed on the sidewall oxide layer 112 and the capping layer 120 by a CVD process as a liner; so that an oxidation of the sidewall of the fin 110 is prevented. A device isolation layer 116 for isolating conductive structures from each other is formed on the silicon nitride layer 114 to a predetermined thickness.
Referring to FIG. 5, the silicon nitride layer 114 above a top surface of the device isolation layer 116 is etched away using a phosphoric acid (H3PO4) solution as an etchant, thereby forming a silicon nitride pattern 114a of which a top surface is coplanar with the top surface of the device isolation layer 116.
When the silicon nitride layer 114 is etched away, the capping layer 120 is also etched away by the phosphoric acid (H3PO4) solution since the capping layer 120 is also comprised of silicon nitride. An optimization of the etching conditions may minimize an etching amount of the capping layer 120. However, the capping layer 120 is inevitably etched away by a certain amount despite the optimization of the etching conditions. Accordingly, the capping layer is somewhat damaged and an original shape of the capping layer is somewhat changed as shown by the damaged capping layer 120a shown in FIGS. 3 and 5. Moreover, when the capping layer 120 is over-etched away, the capping layer is completely removed from the fin 110.
Then, the sidewall oxide layer 112 above the top surface of the device isolation layer 116 is etched away using an aqueous hydrogen fluoride (HF) solution as an etchant, thereby forming a sidewall oxide pattern 112a of which a top surface is also coplanar with the top surface of the device isolation layer 116. As a result, a top portion 110a of the fin 110 is exposed by the sequential etching processes against the silicon nitride layer 114 and the sidewall oxide layer 112.
Referring to FIG. 6, a gate electrode 130 is formed on the device isolation layer 116 in the second direction perpendicular to the first direction, so that the exposed top portion 110a of the fin 110 is partially covered with the gate electrode 130. A dry etching process is performed during the etching process for forming the gate electrode 130, and the damaged capping layer 120a is further damaged and the shape thereof is further changed. In FIG. 6, reference numeral 120b designates a capping layer consecutively damaged by the above etching processes.
As a result, most of the capping layer 120 is damaged and an original shape thereof is changed by various etching processes, and moreover, may be completely removed from the fin 110 after the etching process.
A vertical channel fin FET has been suggested for solving the above problems as disclosed in U.S. Patent Publication No. 2005-145932. According to the above U.S. patent publication, a buffer oxide layer is formed on a nitride layer, which operates as an etch stop layer formed on the fin of the vertical channel fin FET, by a CVD process for preventing the nitride layer from being etched away in a subsequent process. As a result, the nitride layer for the etching stop layer is prevented from being etched away when a nitride layer for a liner is formed in a subsequent process.
However, the above vertical channel fin FET has a problem in that sidewall defects of the fin caused by an etching process for the fin are hardly cured because the buffer oxide layer is formed by a CVD process. The above non-cured sidewall defects of the fin may increase a charge trap on a gate insulation layer and deteriorate the quality of the gate insulation layer, thereby generating current leakage from the channel region. Moreover, when the defects are formed in the channel region, carrier mobility in the channel region is remarkably reduced, thereby decreasing an operation current of the fin FET.